Girl Power? Investment Trends for Female Entrepreneurs on Shark Tank (US)

Final Report
Data Science 1 with R (STAT 301-1)

Author

Allison Kane

Published

December 6, 2023

Introduction

Throughout the corporate sector, female entrepreneurs tend to face greater barriers and often have their businesses undervalued and underrepresented. Many female entrepreneurs find it difficult to break into male-dominated industries, to obtain similar investments to their male counterparts, and to receive the same respect as male entrepreneurs. Additionally, industries with greater representations of women are often viewed as less lucrative, less serious and less deserving of investment.

The reality television show Shark Tank (US) broadcasts real-time deals being made between entrepreneurs pitching their businesses to a panel of investors, called sharks. The show, having run for over a decade, provides a glimpse into the business world for viewers and a wealth of information about how entrepreneurs– both male and female– are treated by investors.

In this EDA, I intend to find if there are differences between the investment and pitching trends of female entrepreneurs, male entrepreneurs, and sharks by analyzing the dataset shark_tank_us_data1. Specifically, I aim to determine if female entrepreneurs are represented as often as male entrepreneurs in the show and which industries featured on the show tend to attract more female entrepreneurs (female-dominated industries). Additionally, I intend to determine if female entrepreneurs ask for similar amounts of investment, valuation, and equity in their businesses as male entrepreneurs, if female entrepreneurs receive investment as often as male entrepreneurs, and if that investment aligns with their initial demands (i.e. quality of investment). Finally, I intend to determine which sharks invest in female entrepreneurs more often and which sharks tend to invest more in female-dominated industries.

Data overview & quality

The shark_tank_us_data is a 285 KB csv file with 1,274 observations and 50 variables, 31 of which are numeric, 16 of which are logicals and 11 of which are characters. A concise high level overview of the dataset(s) being explored. There were some instances of missingness, although some inherent to the data collection process. Some instances of missingness were indications of poor data collection, however most of these missingness issues occurred for variables that were not relevant to the analysis.

Demographics

Figure 1: Gender Ratios

Women are often underrepresented in positions of authority, especially as entrepreneurs. In Figure 1 Shark Tank (US) seasons 1-14, it is clear that women-led businesses are outnumbered by male-led businesses. Mixed gender entrepreneur teams are represented even less. This gender imbalance might limit viewers’ exposure to female entrepreneurs and hinder their acceptance of women in positions of power, especially in business.

Figure 2: Businesses Represented

In Figure 2, the industries most represented include Food/Beverage, Fashion/Beauty, and Lifestyle/Home. Some of these businesses are typically associated with women, like Fashion/Beauty, but representation in leadership roles is rarely dominated by women.

Figure 3: Industries Represented with Female Entrepreneurs

In Figure 3, Industries with female entrepreneurs represented on shark tank (US) seasons 1-14 show that women are most involved in Food/Beverage, Fashion/Beauty, Children/Education and Lifestyle/Home industries. Three industries do not have any female-run businesses represented: Automotive, Electronics, Liquor/Alcohol.

Figure 4: Industry Leadership Gender Representation

Figure 4 indicates that only one industry on Shark Tank (US) has more female-led businesses represented than male or mixed team led businesses: children/education. All other industries have a greater representation of male entrepreneurs than female entrepeneurs and mixed teams. Representation of female entrepeneurs is better in Fashion/Beauty, Health/Wellness and Pet Products industries, but male-entrepeneurs are still represented more often. Representation of female entrepreneurs is the worst in the Automotive, Electronics and Liquor/Alcohol industries. In particular, Liquor/Alcohol does not have any female-led or mixed-team led businesses represented at all.

The bias in gender-representation in Shark Tank (US) could have additional effects, including how frequently deals are made for different genders, the quality of the deal, and how individual sharks behave towards different genders of entrepeneurs.

Conclusions

According to this analysis, there are some differences in the experience of female and male entrepreneurs on Shark Tank (US) seasons 1-14. Female entrepreneurs are underrepresented on Shark Tank (US) and are underrepresented in most industries, only outnumbering male entrepreneurs in the Children/Education industry. Some industries lack completely in female entrepreneurs, including the Automotive, Electronics, and Liquor/Alcohol industries (see Figure 1, Figure 2, Figure 3 and Figure 4). This was expected, as female entrepreneurs are generally underrepresented across most industries outside of Shark Tank (US) and many industries that are historically aligned with STEM fields, like electronics and automotive industries, are overwhelmingly male dominated.

Deals on shark tank are made approximately 60% of the time, with female entrepreneurs and mixed gender teams making deals more frequently and male entrepreneurs less frequently (see Figure 5 and Figure 6). This was somewhat unexpected, as I anticipated female entrepreneurs might make deals less often with expected less investment from sharks. However, this difference could be attributed to male entrepreneurs turning down deals more frequently or female entrepreneurs being less likely to counter initial offers form sharks. Follow up studies could be done to determine if this is the case. Additional analyses showed that some industries with strong representation of female entrepreneurs made deals more frequently than other industries. However, the automotive industry–an industry with no female entrepreneurs represented–made deals overwhelmingly the most often (see Figure 7).

Analysis of female, male and mixed gender entrepreneurs’ initial expectations for investment, equity and valuations and their deal values showed that male entrepreneurs typically overvalue their businesses. Female entrepreneurs typically ask for and receive less investment and offer and give more equity (see Figure 8, Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14, Figure 15, Figure 16, Figure 17, Figure 18, and Figure 19). Female entrepreneurs also typically left pitches with valuations that were more in line with their initial expectations than their male or mixed gender team counterparts. This could indicate that female entrepreneurs have a more accurate understanding of their business’s worth than their male or mixed gender team counterparts.

Investments received were typically in line with initial expectations across all genders, while entrepreneurs typically gave slightly more equity than expected (see fig-20, fig-21, and fig-22). This was somewhat unexpected, as I anticipated that there might be a difference across gender how in line a deal was with initial entrepreneur expectations, perhaps with female entrepreneurs not receiving quality deals. The differences in overall investment amounts and equity amounts seen in previous figures might be a case where female entrepreneurs are asking for less investment and offering more equity than their male counterparts and, as a result, receive offers reflecting their initial expectations. Female entrepreneurs also might go about negotiations differently, driving these discrepencies across genders. Follow-up studies could be conducted to determine how often female entrepreneurs counteroffer, walk away from deals, or use sharks as leverage to enhance another deal. Additional data would be needed to conduct these studies.

In analysis of shark behavior, male and female sharks typically made deals at similar rates (most less than 10%). However, Lori Greiner and Mark Cuban had higher rates of deals than most other sharks (15.6% and 18.1% respectively), which was somewhat unexpected. In making deals with female entrepreneurs, male sharks were more likely to value businesses less than their ideal valuation compared to their male counterparts (see Figure 34). Additionally, in their interactions with female entrepreneurs, it was clear that female sharks invested in female-led businesses much more often than male entrepreneurs, even despite differences in individual shark investment frequencies (see Figure 36). Very minor differences were seen in how often male sharks invest in women-dominated industries compared to female sharks, with male sharks favoring Fitness/Sports/Outdoors over Children/Education focused businesses. This is somewhat unsurprising, as generally male and female sharks have similar industry preferences. However, Children/Education might appeal more to female sharks, or female entrepreneurs in those industries might prefer to work with female sharks more than male sharks.

Moving forward, additional studies must be done to better understand this data. Additional data sources could be joined to help identify how often sharks counteroffer, walk away from offers, or encourage bidding wars. Also, further analysis could be conducted of how treatment of female entrepreneurs has changed over time. Shark Tank (US) has been on the air since 2009, indicating that it could be beneficial to see how representation of women, investment/equity/valuation trends, shark behavior and industry demographics have changes in its seasons. Additional seasons of shark tank could be added to this dataset to enhance this analysis.

References

Thirumani, S., Rehman, A.U., and Molagoda, J., 2023, Shark Tank US dataset, https://www.kaggle.com/datasets/thirumani/shark-tank-us-dataset

Appendix: Technical Info

Some instances of missingness were critical for the data collection process. For example, in instances where deals were not made, some missingness was present in variables like Total Deal Amount (reflecting that there was no deal made). In the cleaned data, these values were adjusted to limit missingness, often replacing NA values with 0.

There are seven instances where the pitcher’s gender is not included in the dataset. This might be an instance where the pitcher did not identify with a binary option of gender (might identify as non-binary, genderqueer, etc.). It could also be an instance where their gender identity was not collected. Given that there are only seven of these observations, there could be limited analysis on this population can be made and, therefore, are excluded from the cleaned dataset.

Other instances of missingness in variables like Company website, or Pitchers State did not pose significant issues to the analysis, as they are not as relevant.

Logicals also had instances of using numbers 1 and 0 to represent TRUE and FALSE. These values were adjusted in the cleaned data to be either TRUE or FALSE.